Best sounding LAN cable for DS

General HiFi discussion, using the Tune Method to evaluate performance

Moderator: Staff

Post Reply
User avatar
sommerfee
Very active member
Very active member
Posts: 337
Joined: 2007-02-02 17:40
Contact:

Best sounding LAN cable for DS

Post by sommerfee » 2008-03-13 22:05

I bought some more no-name patch cables from the electronic shop just around the corner, so in total I have now 7 cables as possible connection between WLAN router (a Netgear FWG114v2) and DS. These are the cables:

1. 1.5m Cat5 UTP 24AWG (was shipped with the Netgear router, I have used this as connection so far)
2. 2m Cat5e S-STP 26AWG (new, no-name)
3. 2m Cat5 SFTP 26AWG (old & used, from "lanpro")
4. 3m Cat5 FTP 26AWG (very little used, no-name)
5. 3m Cat5e SFTP 26AWG (old & used, from "e'quip")
6. 5m Cat5e SFTP 26AWG (new, no-name)
7. 10m Cat5 FTP 26AWG (little used, no-name)

No.1 is definitely the worst of them. It sounds "cloudy" and "undecided"/"non-rocking", so I planned to complain about my unit after some weeks of burn-in and not getting better.

No.2 and 3 are the best ones and quite on par regarding tune-dem. No.2 sounds a little bit less colorful than No.3, but maybe this is because cable No.2 is brand new?

No. 4 and 7 where quite on par, and worse than 2+3.

Since until now "short" was better than "longer" I expected that No.6 would be worse than No.5, but is was the opposite. Funny. And both (5+6) seems to be better than 4+7.

Conclusion:
It seems that short cabled tend to be better than longer ones. (But not always.) Furthermore good shielded ones (SFTP) seems to be better than less shielded ones (FTP), and unshielded (UTP) seems to be the worst. Shielding seems to be more important than the length of the cable.

I hope that more tests will give more light to the situation. Paolo, what cables have you tried so far and what are your conclusions so far?

Open questions:
Are the above conclusions valid? Do the cables need a "burn-in"? Does the direction matter? And of course: WHY?

Future:
At first I will keep No.2 and make another comparison with No. 3 in a week, to check if burn-in is relevant. Furthermore I try to get even better (shielded) cables, like CAT6 ones.

Boy, I really hope that Linn will offer an own patch cable very soon...

Axel

P.S.: See http://en.wikipedia.org/wiki/Twisted_pair about shielding
Last edited by sommerfee on 2008-03-19 18:49, edited 3 times in total.

User avatar
Music Lover
Very active member
Very active member
Posts: 1667
Joined: 2007-01-31 20:35
Location: In front of Lejonklou/JBL/Ofil

Post by Music Lover » 2008-03-13 22:18

WOW :shock:
Can you also connect DS directly to the PC, avoiding the router and all other connections (i.e. internet)?
It may sound better...
Please test.
It's all about musical understanding!

JohnS
Member
Member
Posts: 48
Joined: 2007-10-02 21:59

Post by JohnS » 2008-03-13 22:58

In the next week I'm going to try a number of configurations and cable combinations:

Config 1: This is basically a DS connected directly to the Server.
- Home Router connected to the Music Server (network interface 1)
- Music Server (network interface 2) direct to DS with a crossover cable
- Bridged connection on the server (for DHCP and Cidero Control)

Config 2: This is with a dedicated switch between the Music Server and DS. Localised traffic between the two.
- DS connected to dedicated switch
- Dedicated switch to DS
- switch uplink to home router (for DHCP and Cidero control)

I'll try shielded and unshielded cables for one or all of the connections. From what I hear so far I am guessing a short shielded connection going into the DS will make a difference, but does the whole network need to be shielded? Will the switch introduce interference or create a clean signal?

Being ethernet I have a hard time believing that data transfer is compromised. Ethernet goes up to 100m on 1Gps data networks even with Cat5e unshielded cable in a complex office network without problems, and this is only 100Mbps over a few metres - so no problem at a data level.

I speculate that what we must be talking about is the DS being rather sensitive to RFI or other signals on the ethernet cables. I wonder if Linn engineers have totally isolated the digital to analogue circuitry? speculation again.

I'm not 100% sure of cable types, I've seen :

Unshielded : UTP
Shielded : FTP, SFTP, SCTP, SSTP

Anyone know the difference between the shielded types??

Thanks
John

JohnS
Member
Member
Posts: 48
Joined: 2007-10-02 21:59

Post by JohnS » 2008-03-13 23:00

Sorry - I shouldn't have missed a simple typo......

Config 2: This is with a dedicated switch between the Music Server and DS. Localised traffic between the two.
- DS connected to dedicated switch
- Dedicated switch to Music Server
- switch uplink to home router (for DHCP and Cidero control)

User avatar
sommerfee
Very active member
Very active member
Posts: 337
Joined: 2007-02-02 17:40
Contact:

Post by sommerfee » 2008-03-13 23:24

Music Lover wrote:Can you also connect DS directly to the PC, avoiding the router and all other connections (i.e. internet)?
It may sound better...
Please test.
I'm sorry, but this is not easy to test for me. At first the Linn applications do not run correctly on the PC which plays the role as Twonky Media server. They find the DS, but the configuration application always shows the same rubbish data ("display intensity 0" etc.) and the GUI application makes trouble as well.

The other problem is that I do not have a display for the server PC, so I use VNC, having my Nokia as VNC viewer. So I need WLAN.

Axel

JohnS
Member
Member
Posts: 48
Joined: 2007-10-02 21:59

Post by JohnS » 2008-03-13 23:40

Hi Axel - That's strange that it doesn't work. I've been trying a few things, and currently I have a laptop running XP Pro, Twonky, Cidero and the Linn config programs if I want them.

It's connected directly to the DS with a crossover adapter and it all works. I had to set a static IP address and run a DHCP Server on the laptop for the DS to get its address.

I only have UTP cables right now and I'm just waiting for an SSTP cable and a new switch to arrive in the post.

FYI: the cables I ordered can be seen on amazon.co.uk search for SSTP they are by a company called Lindy.

DigitalSources
New member
New member
Posts: 5
Joined: 2008-03-11 23:51
Location: Germany
Contact:

Post by DigitalSources » 2008-03-13 23:41

Axel,

that sounds interesting!
I will try the same over the weekend.

There is ne thing that I think you should test as well: The influence of the connection between the Router/Switch and the PC/NAS. Is there also a sound difference between the various cables if they are used to connect the PC to your Switch?

By the way: My setup is more complex. I have a Router (FritzBox) that connects to two Switches. On one of the switches there is the PC with Twonky and on the other switch there is the DS connected. So a music signel has to pass three cables and three network devices to reach the DS. Let's see what results I might produce.

Best regards,
Andreas.

www.digitalsources.de

User avatar
sommerfee
Very active member
Very active member
Posts: 337
Joined: 2007-02-02 17:40
Contact:

Post by sommerfee » 2008-03-14 12:08

What I forgot to mention was the amount of differences: Between cables 2-7 there were only minor differences, they remind me of the differences between one to three Unidisk options, e.g. digital output on/off. So nothing spectacular, but IMHO definitely worth trying and confusing. But what shocked my dealer and me was the amount of difference of cable No.1 to the other ones. (about 5 relevant Unidisk settings?) Was it because this cable is unshielded? Was it because it was an 24AWG cable (thinner than 26AWG)? Was it because this cable is blue while the other ones are grey resp. yellow? :lol:

@JohnS: Yes, that's strange, since all other network applications work well on this PC, and LinnConfig and LinnGUI work without problems on two other PCs here, connected to the DS via LAN or WLAN.

I'm curious about your experience with the new cable, please report after listening.

@Andreas: Will test this today evening and report. BTW: I also tried to add another switch (Netgear FS105) between router and DS, and this has made the sound a little bit worse.
Last edited by sommerfee on 2008-03-15 08:09, edited 1 time in total.

User avatar
lejonklou
Administrator
Administrator
Posts: 4570
Joined: 2007-01-30 10:38
Location: Sweden
Contact:

Post by lejonklou » 2008-03-14 13:34

Thanks for starting this topic, Axel, and for facing these questions with an open mind and a methodical trial-and-error approach! This is the way to do it IMO. By sharing information, we could rather quickly end up with a recommended list of DS peripheral parts, which I think would be a great achievement!

A note to anyone wanting to join this search is that the methodology is crucial. The comparison between parts needs to be done systematically, with no other system changes that could affect the outcome. It also needs to be done with a critical mind - with critical I mean reproducing results several times and making sure it's the change you are making that is causing the change in the music.

Of course it needs to be done with the Tune Method, but I hope that's obvious to everyone. So far I have found that those who don't use the Tune Method seem to think that the differences obtained by changing parameters in the digital domain are negligable or nonexistent. This doesn't surprise me, because in my ears, the sound is close to identical while the tune and the way I perceive the musical flow can be very clearly affected by, for example, a change of digital cable.

We don't yet know whether the results we get can vary between different hardware, but my suspicion is that the large differences we get are more likely to be universal. For example, a certain cable type that sounds really bad will most likely sound really bad in all systems. The length of cables, however, could very well vary between different systems OR in reality depend on something else - like the individual connectors or the connection between cable and connector. The smaller the difference, the higher the number of tests you need to make to be certain and the more samples you need to make sure you're actually comparing the parameter you think.

By the way; 24 awg is thicker than 26 awg. Not the other way around.

Keep up the good work! This kind of "knowledge building" excites me and was exactly what I had in mind when this forum was created.

paolo
Active member
Active member
Posts: 124
Joined: 2007-01-31 12:49
Location: Rome, Italy

Re: Akurate/Klimax DS LAN cable hell

Post by paolo » 2008-03-14 18:58

Sommerfee, thank you for your extensive report, i find it interesting and very useful!

I'm actually try comparing different switches to decide which is for me the best "architectural" solution with the DS, so I've still not made any methodic comparison with network cables. I've Just tried some I had laying around and left working the one I thought was better. I agree the difference is musically relevant, surprisingly so. I've also quickly compared the two directions of a cable, it doesn't seem to make great difference, but again I'll not be 100% sure until I do a serious test.

About your problems with the Linn Gui, maybe it's s a trivial qs., anyway have you installed Microsoft .NET 2.0?

DigitalSources, yes in my experience also the network cable between the library (NAS or PC) and the switch is clearly audible. Given the fact that any switch I've tried has had an influence on the sound of the DS, I suggest you to keep the number of network devices in the path between the library and the DS as low as possible (if feasible try to connect DS and library to the same switch).
JohnS wrote:Being ethernet I have a hard time believing that data transfer is compromised. Ethernet goes up to 100m on 1Gps data networks even with Cat5e unshielded cable in a complex office network without problems, and this is only 100Mbps over a few metres - so no problem at a data level.

I speculate that what we must be talking about is the DS being rather sensitive to RFI or other signals on the ethernet cables.
Yes, I agree. I can add that the "phase" on the mains is clearly audible both with the NAS and the switch, just like with any component in the audio system. This reinforces to me the idea that there must be an "analogical" interaction between the network devices and the DS through the network cable. I cannot think of any way mains phase can affect the transmission of data packets infact.

Lejonklou wrote: So far I have found that those who don't use the Tune Method seem to think that the differences obtained by changing parameters in the digital domain are negligable or nonexistent. This doesn't surprise me, because in my ears, the sound is close to identical while the tune and the way I perceive the musical flow can be very clearly affected by, for example, a change of digital cable.
Fredrik, I completely agree with any word you've wrote here! This is still more evident in my experience when comparing things when the possible analogical interactions are completely excluded, for instance the sound of the same music ripped with different ripping SW or the same SW with different settings. Here the sound is very often really *absolutely* the same. But still very often, using tune dem, there are differences in the way the music communicates and can be followed. Really interesting!

Quick report about switches: I've tried also a Netgear FS608 FastEthernet switch, I was curious to know if there was connection between the poor "sound" of the Gigabit brother GS608 and the speed of switch: no, the FS sounds very similar. It is probably the way they're build that makes them "sound" a bit uncohesive weak in the bass.

Paolo

ThomasOK
Very active member
Very active member
Posts: 2661
Joined: 2007-02-02 18:41
Location: United States
Contact:

Post by ThomasOK » 2008-03-15 00:26

Thanks for all the trials Axel, having done a similar thing just recently with the AC cables I know how much work is involved. It is interesting that the one thicker cable sounded worse although since it was also unshielded it is hard to say where the problem lies. It does seem that many of the thicker AC cables aren't as good as the more common thickness so gauge obviously has nothing to do with quality.

I would be really interested to find out what you would find using a 4 meter cable. Although I have not tried this the Nordost people claim that cables sound best at 4 meters. They say that it doesn't matter whether it is an interconnect, a speaker cable or a power cable they all sound best at 4 meters. They also say they have no idea why it is it is just what they have found and that it seems to hold regardless of brand or construction technique. One of my co-workers just came back from a Nordost training and they demonstrated the same power cable at a number of different lengths and he did indeed feel that the 4 meters was somewhat better. Just another one of those wierd cable things I thought might be worth trying. Sounds crazy, but who knows!

User avatar
sommerfee
Very active member
Very active member
Posts: 337
Joined: 2007-02-02 17:40
Contact:

Post by sommerfee » 2008-03-15 08:23

I can confirm what paolo has found out: Unfortunately the cable between PC/NAS and router makes a difference, too. When I used the "bad blue" cable between them, a similar effect occurred. I did not test this out, but my feeling was that although the effect was similar, it was not as worse as when this cable was used between router and DS. (Will check this later.)

@paolo: Yes, it's running Windows XP-SP2 with .NET 2.0 installed, otherwise I could not start the applications without getting an error. But they pretend working, but show always the same wrong data. The PC has no extra software on it (and never had) except Twonky Media, VNC, and the Linn apps. And it wasn't working before I installed Twonky and VNC either.

Axel

paolo
Active member
Active member
Posts: 124
Joined: 2007-01-31 12:49
Location: Rome, Italy

Re: Akurate/Klimax DS LAN cable hell

Post by paolo » 2008-03-16 12:13

I've also quickly compared the two directions of a cable, it doesn't seem to make great difference, but again I'll not be 100% sure until I do a serious test.
I was wrong guys! Just tried today, there's definitely a difference with the two directions of the same network cable. The difference is not negligible.

I've bought also a couple more ftp cables and I can confirm what Axel has found: I have 3 ftp cabes now and they all sound better than any utp cable I've tried so far. Seems it could become a good rule, shielding is better! I've still not tried any SFTP/SSTP at now.

Paolo

Charlie1
Very active member
Very active member
Posts: 2992
Joined: 2007-12-11 00:30
Location: UK

Post by Charlie1 » 2008-03-16 13:13

Hi Paolo, sounds like you're all coming to the same conclusions regarding the specification and direction of Ethernet cables.
On the subject of simplifying the signal path, I know you've found a crossover better, but have you also found that multiple devices degrade Tune Dem - for exampe: DS to Switch, then router, then another switch and finally the PC etc?

User avatar
sommerfee
Very active member
Very active member
Posts: 337
Joined: 2007-02-02 17:40
Contact:

Post by sommerfee » 2008-03-16 15:16

sommerfee wrote: @paolo: Yes, it's running Windows XP-SP2 with .NET 2.0 installed, otherwise I could not start the applications without getting an error. But they pretend working, but show always the same wrong data.
Addendum: The new LinnGUI 1.2-0.9 is running fine on this PC, so I assume that Linn has already found the problem and fixed it. (Hopefully they will release a new LinnConfig with this fix, too.)

JohnS
Member
Member
Posts: 48
Joined: 2007-10-02 21:59

Post by JohnS » 2008-03-17 14:31

I'm still waiting for my SSTP cables, but my new Gigabit switch arrived, a Netgear GS105.

I've only listened for an hour with UTP cables, but I would say it's a bit better than a crossover cable direct to my Dell laptop running Twonky. Could the quality of the network driver be better from one device to the other, more noise from the Dell? Are we going to be comparing Network Interface Card's (NIC)? It'll be interesting to see with shielded cables later in the week.

The switch is powered by an external 'brick' power supply providing DC +/- into the switch. It has an earthing point, should I connect this to the DS?

John

User avatar
Music Lover
Very active member
Very active member
Posts: 1667
Joined: 2007-01-31 20:35
Location: In front of Lejonklou/JBL/Ofil

Post by Music Lover » 2008-04-02 16:26

Gents, any update on your testing activities?
- what cable is best/worse
- what switch is best/worse
- what system architecture is best/worse
It's all about musical understanding!

User avatar
Music Lover
Very active member
Very active member
Posts: 1667
Joined: 2007-01-31 20:35
Location: In front of Lejonklou/JBL/Ofil

Post by Music Lover » 2008-04-02 18:01

Did a search on the Net and found cat6 cables in two models; FTP and SSTP.
Anyone tried these?
Last edited by Music Lover on 2008-04-02 18:13, edited 1 time in total.
It's all about musical understanding!

User avatar
sommerfee
Very active member
Very active member
Posts: 337
Joined: 2007-02-02 17:40
Contact:

Post by sommerfee » 2008-04-02 18:12

I have ordered some more cables (cat.6 included) and will do further tests as soon as they arrive. Furthermore I will do a comparison between three different switches on this week-end.

User avatar
Music Lover
Very active member
Very active member
Posts: 1667
Joined: 2007-01-31 20:35
Location: In front of Lejonklou/JBL/Ofil

Post by Music Lover » 2008-04-02 18:26

sommerfee wrote:I have ordered some more cables (cat.6 included) and will do further tests as soon as they arrive. Furthermore I will do a comparison between three different switches on this week-end.
Have fun!
Do you have the names of the switches?
It's all about musical understanding!

JohnS
Member
Member
Posts: 48
Joined: 2007-10-02 21:59

Post by JohnS » 2008-04-02 21:31

I’ve been trying a few combinations of cables and configurations to get the best of the DS.

UTP, FTP and SSTP cables in various lengths and a few network combinations. I can’t try all the combinations as there are just too many.

I dedicated a Vista PC running Twonky with local music files as my ‘Server’ – also importantly it has two network cards that have been bridged so in effect the machine can effectively act as a switch, that way I can connect one to my general network and the other directly to the DS, or just use one and use the server as a machine on the network . When bridged the DS can access the general network and home router DHCP server through the server and get and IP address and access the controller (no fixed IP’s or software DHCP servers).

All my configs allow a wireless Portable PC running Cidero to access the DS and visa-versa.

My home wireless router (Netgear DG834 v3) manages a wireless network, a skype phone, various laptops etc. Busy!

Config 1
  • 5m UTP to DS -> home wireless router -> 5m UTP to Server
    Optional : Put a switch just before the DS and a short shielded cable.
    Pros: easy and typical config
    Cons : DS Traffic mix with other network traffic, quality of home router?
Config 2
  • Server -> 5 port gigabit switch (Netgear GS105) -> DS
    Home wireless router -> 5 Port gigabit switch (GS105 as above - i.e. a 3 way start from the switch)
    Pros: DS traffic localised by switch
    Con’s : another switch
Config 3
  • Server -> DS (crossed or straight - both work without any extra crossover plugs)
    Server -> Home wireless router
    Optional : try swapping the two over to see if one port is better than the other.
    Pro’s : DS traffic isolated, not even a switch between server and DS
    Cons : need server near to the DS
What I’ve found is that config 1 is clearly the worst, although interestingly, when a put in the optional switch with a 0.5m SSTP cable just before the DS, it improved significantly. I can only suggest that it ‘cleaned’ the signal by receiving the data packets and re-transmitting cleanly, it wasn’t however up with the best which suggests that going via the home router was a bad move.

However Config 2 and 3 are clearly better in terms of Tune Dem, and are very close between the two. So far I think I slightly prefer config 3.

As for cables, in config 2 and 3, I only found a small difference in the cables, but I’ve only tried short runs. I did try a 5m and a 2m UTP cable against a 3m FTP and 2m SSTP directly between the server and DS. The 5m UTP cable has a slight effect, the 2m UTP and the two shielded cables were the best and almost identical, although I need to listen a bit more to be sure.

I also tried getting some music playing and unplugging the cable to the general home network and there was no difference, so no interference or impact.

I noticed that in option 3, there was a minor difference between which port was connected to the DS, one was a motherboard port and the other an Intel PCI card. I think the Intel card is slightly better, but again it’s hard to say as the effect is minimal.

So my conclusions so far are that a ‘dedicated’ network between the DS and Server is important, this can be done by direct cabling or use of a network switch to localise traffic. Shielded cables do seem to make a difference, but minimal for short runs. I’ll probably stick with config 3 with a shielded cable for now.

Json
Member
Member
Posts: 13
Joined: 2008-01-26 23:05

What about grounding?

Post by Json » 2008-04-03 14:55

Hi all!

This cable issue is really interesting!
I have some questions for you who have tried different cable types, especially shielded cables.

1) Where do you ground the cables?
- I suppose the DS has a shielded ethernet connector so could it be used for grounding? (via the power cable)
- Most switches/routers I have seen seem to use un-grounded power cables that would not make it possible to ground the patch cable from the router? (I have not seen any "ground points" on the routers either, but maybe I haven't looked hard enough).
- What about grounding through a computer connected to the same switch/router? (again via a grounded power cable)
- Or the NAS?

2) I read somewhere (can't find the link now unfortunately) that the network should only be grounded in one place.
- If so, how do you avoid grounding in more places? For example, if the DS grounds the network and the NAS grounds the network (since both use grounded power cables).
- Some people seem to have removed the shield from one end of the cable, and if so what would be the best way to do that? Remove the metal shield on the end plug of the patch cable, or somehow cut the ground wire inside the cable?
- In such a setup should the DS be the grounding point, or should it be something else?

JohnS
Member
Member
Posts: 48
Joined: 2007-10-02 21:59

Post by JohnS » 2008-04-03 15:56

That's a great point about grounding. My Netgear GS105 has metal body and a ground point, but I've not connected it to the DS which also has a ground point. I hesitated about connecting the two, but then what about the computer, should I connect all three? The shielded cables all have metal bodies to the plugs, so I was hoping they would connect anyway.

Charlie1
Very active member
Very active member
Posts: 2992
Joined: 2007-12-11 00:30
Location: UK

Post by Charlie1 » 2008-04-03 15:56

Hi Json,
I'm really only familiar with Cisco equipment these days, but I do recall a product supplied with Ethernet cables that had a metal clip on only one end, located approx 1ft from the connector. This would then be screwed to the chassis.

Nowadays the actual connector of an FTP Ethernet cable is partly metal and the Cisco's are earthed within the actual RJ45 socket. Cisco state the following warning regarding earthing by the way, "Category 5e and Category 6 cables can store high levels of static electricity because of the dielectric properties of the materials used in their construction. Always ground the cables (especially in new cable runs) to a suitable and safe earth ground before connecting them to the module."

Cisco also require grounding each end "To comply with GR-1089 intrabuilding, lightning-immunity requirements, you must use foil-twisted pair (FTP) cable that is properly grounded at both ends", but I don't know what's changed so that they don't need to worry about earthing loops anymore.

I presume the RJ-45 cable supplied with the DS products is at least FTP spec and the RJ-45 jack is earthed?

User avatar
vicdiaz
Active member
Active member
Posts: 228
Joined: 2007-02-06 04:37
Location: Trujillo Alto, PR U.S.A.
Contact:

Post by vicdiaz » 2008-04-03 18:49

Charlie1 wrote:I presume the RJ-45 cable supplied with the DS products is at least FTP spec and the RJ-45 jack is earthed?
:shock:

What RJ-45 cable???? There was no Ethernet cable supplied in my Akurate DS box, only the serial cable to connect the DS to the Kontrol.
Vic
Ivor's "Tune-Method Seminar" Alumni

Post Reply